The Number Factory

For Lindall and Olson, peril lies in seeing their products misused and the results discredited. They emphasize training to keep clients from unintentionally tipping the scales as they weigh economic questions.

For the rest of us, peril lies in not understanding how those scales can be tipped—unintentionally or on purpose—in producing the numbers that fuel public debate. So here, a safeguard: a look inside Minnesota Implan Group, the place where “multipliers” get made.

Input-Output 101

First, a little of the economic theory behind Implan.

Implan is built on a mathematical input-output (I-O) model developed by Wassily Leontief, the 1973 Nobel laureate in economics, to express relationships between sectors of the economy in a chosen geographic location. In expressing the flow of dollars through a regional economy, the input-output model assumes fixed relationships between producers and their suppliers based on demand. It also omits any dollars spent outside of the regional economy—say, by producers who import raw goods from another area, or by employees who commute and do their household spending elsewhere.

The idea behind input-output modeling is that it’s the interindustry relationships within a region that largely determine how that economy will respond to change. In an I-O model, the increase in demand for a certain product or service causes a multiplier effect, layers of effect that come in a chain reaction. Increased demand for a product affects the producer of the product, the producer’s employees, the producer’s suppliers, the supplier’s employees, and so on, ultimately generating a total effect in the economy that is greater than the initial change in demand.

“Say demand for Andersen Windows’ wood window products increases,” Lindall says. “Sales grow, so Andersen has to hire more people, and [the company] may buy more from local vendors, and those vendors in turn have to hire more people . . . who in turn buy [more] groceries.”

The ratio of that overall effect to the initial change is called a regional multiplier, and it’s sometimes expressed like this:

(Direct Effect + Indirect Effect) / (Direct Effect) = Multiplier

More often, though, that added element of household spending on things like groceries is included, and then the ratio of overall effect to original change is expressed like this:

(Direct Effect + Indirect Effect + Induced Effect)/(Direct Effect) = Multiplier

The term “multiplier” can’t be used generically, though. Multipliers always express the ratio of overall effect to initial change by one of three measures: output (dollars’ worth of production), labor income, or jobs generated. So there are output multipliers, income multipliers, and employment multipliers.

In gauging economic effects, “you always have to ask what kind of multiplier are you talking about,” Lindall says. Users of Implan or of any input-output model “have to make the determination of what the initial change in the economy is,” he explains—that is, whether they want to study a change in production, labor income, or employment.

He returns to his earlier illustration using Andersen Windows, a case involving an output multiplier that would have been calculated using data specific to the window manufacturing industry and specific to the county where Andersen is based: “So for example, there’s a million dollars in new production of windows in Washington County. That million dollars of new output is multiplied by the multiplier for the window manufacturer, and it may be, like, 1.2. Then the total impact on the economy is $1.2 million. So there’s an additional $200,000 worth of activity in Washington County because of that new production that was occurring.”